<p dir="ltr"><br>
On Oct 21, 2013 6:20 PM, "Dave Taht" <<a href="mailto:dave.taht@bufferbloat.net">dave.taht@bufferbloat.net</a>> wrote:<br>
><br>
> Having more accurate dropped information in this qdisc is useful.<br>
><br>
> Signed-off-by: Dave Taht <<a href="mailto:dave.taht@bufferbloat.net">dave.taht@bufferbloat.net</a>><br>
> ---<br>
> net/sched/sch_fq_codel.c | 1 -<br>
> 1 file changed, 1 deletion(-)<br>
><br>
> diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c<br>
> index 5578628..437bc95 100644<br>
> --- a/net/sched/sch_fq_codel.c<br>
> +++ b/net/sched/sch_fq_codel.c<br>
> @@ -193,7 +193,6 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch)<br>
> list_add_tail(&flow->flowchain, &q->new_flows);<br>
> q->new_flow_count++;<br>
> flow->deficit = q->quantum;<br>
> - flow->dropped = 0;<br>
> }<br>
> if (++sch->q.qlen <= sch->limit)<br>
> return NET_XMIT_SUCCESS;<br>
> --<br>
> 1.7.9.5<br>
I am travelling to Edinburgh, so will be short.</p>
<p dir="ltr">Since fqcodel recycles a slot, we need to clear this counter. We do no know if slot is reused by previous flow or a new flow hashing to same bucket.<br>
</p>